We applied a recently developed integrative clustering method to the Cancer Genome Atlas (TCGA) pan- cancer dataset and identified a novel luminal-TP53 breast cancer subtype. Luminal breast cancer typically expresses estrogen receptor (ER) and/or progesterone receptor (PR), and is associated with low TP53 mutation rate (15-30%) which is a factor linked with poor prognosis. Surprisingly, the luminal-TP53 subtype we discovered, showed substantially more frequent TP53 mutation (>60%), and displayed molecular features strongly resemble those of basal-like breast cancers, which characteristically lack ER, PR and HER2 expression. Reverse-phase protein array data analysis revealed that this subtype displayed higher level of phospho-AKT expression as compared to other breast cancer subtypes. Given that aberrant signaling through the PI3K-AKT signaling pathway has been shown to constitute a mechanism of endocrine resistance, we posit that these tumors are unlikely to respond to hormone therapy but are possible candidates for combinatorial therapies including PI3K/AKT/mTOR inhibition. This proposal aims to develop an integrated genomic signature along with clinical methods to classify the luminal-TP53 subtype with high accuracy, and to validate this novel subtype in independent datasets including the METABRIC cohort and a retrospective MSKCC cohort with long- term follow-up. We will also use cell line and patient-derived xenograft models to study the molecular drivers and therapeutic targets in luminal-TP53 tumors.